Modified Short Proximal Femoral Nail for Intertrochanteric Fractures of Femur in Indian Patients - our Experience

Introduction: Proximal femoral nail (PFN) is a commonly used implant for intertrochanteric fractures which is designed according to western femoral measurements. However, anthropometry of proximal femur in Indian and in general, Asian, are smaller. So a modified short PFN with smaller dimensions was developed. This study analyses the radiological and functional outcome of treatment of intertrochanteric fractures with modified short PFN. Materials and Methods: A retrospective study analysed 120 adult patients operated between 2014-2017 using modified short PFN for intertrochanteric fractures, having a minimum follow-up of 12 months. Clinical and radiological parameters including tip-apex distance (TAD), position of tip of lag screw in femoral head, lateral slide of lag screw as well as length of anti-rotation screw were measured. Final functional outcome was assessed using Barthel’s index and Kyle’s criteria. Results: Good reduction was achieved in 90.83% cases and 79.16% had ideal placement of lag screw in femoral head. Intra-operative difficulties were encountered in 13.33% (n=16). Mean TAD AP (anteroposterior) was 11.8mm, TAD LAT (lateral) was 11.0mm and mean TAD TOT was 22.8mm. Overall mean lateral slide was 3.20mm and it was more in unstable fracture. We had five mechanical failures, one patient with screw breakage without loss of reduction and two peri-implant fractures after union. 81.66% returned to pre-injury levels of activity with 88.33% good to excellent outcome as per Kyle’s criteria. Conclusion: Although, not devoid of complications, modified short PFN results in good functional recovery of patients with intertrochanteric fractures of femur.


INTRODUCTION
Intertrochanteric fracture (IT) of femur is a very commonly encountered orthopaedic condition especially in geriatric population 1 . Conservative management of these fractures are fraught with complications of prolonged recumbency as well as limp and shortening due to malunion in coxa-vara 2 . Extramedullary implants such as DHS, once considered the solution to these fractures have performed less than satisfactorily in unstable patterns paving way for intramedullary implants [3][4][5] . One such intramedullary implant currently in vogue is proximal femoral nail (PFN). Proximal femoral nail (Synthes) was designed keeping in mind the western population and comes in size of 240mm length, 17mm proximal diameter, distal diameter of 10-12mm and sizes of proximal and distal cephalic screws being 11mm and 6.5mm respectively. Study by Su et al demonstrated marked variability in location of femoral isthmus across various ethnic groups 6 . Siwach , in his study on 150 femoral bones, demonstrated smaller measurements related to proximal femoral and isthmus in Indian femurs. He recommended, in order to reduce incidence of intra-operative complications like fractures and splintering, implants need modifications according to Indian anthropometry. He recommended cephalomedullary nail to be adapted to dimensions described by Leung et al 7 . Leung et al used modified gamma nail for use in east Asian population and demonstrated improved clinical results in their multicentric study 8 . Indian femurs are proven to have considerably smaller anthropometric measurements compared to western population, thus requiring a smaller implant 9 . Pathrot and colleagues advised certain modifications in the short proximal femoral nail available in the Indian market 10 . Modified short proximal femoral nail also called trochanteric fixation nail (TFN) was introduced by Yogeshwar implants private limited for the purpose of Indian population and it works on the principles of PFN. Very few studies are published using this implant although it is being used very frequently 11,12 . We undertook this study to assess the clinical and radiological outcomes of intertrochanteric fractures of femur treated with modified short PFN.

MATERIALS AND METHODS
A retrospective study was conducted in patients that were operated at our institutes with modified short proximal femoral nail for IT fractures of femur between January 2014 to December 2017. Inclusion criteria for the study was kept as skeletally mature patient, with fresh (< 2 weeks) intertrochanteric fractures, treated operatively using modified short PFN, having a minimal follow-up of at least 12 months. Exclusion criteria were, fractures extending well below lesser trochanter, associated with other fractures and inadequate medical follow-up records or radiographs.
A thorough search of records were done at the respective institutes of authors (both the institutes are tertiary care referral centres with dedicated trauma centres) and we found that 147 patients were operated with modified short PFN during this period. However 27 records were disqualified mainly due to inadequacies in radiographs, and lack of minimum 12 months follow-up. 120 records qualified for the study. The implant used was similar to PFN (Synthes), manufactured and distributed by Yogeshwar implants private limited (Thane), except it being smaller in size (Fig. 1). The implant is approved for use by Indian FDA. Nails were made of 316L stainless steel. Length of the nail used was 180mm with a proximal diameter of 15mm. Distal diameter had options of 9, 10 and 11mm. Two cephalic screws placed using jig measured 8.0mm (lag screw/lower screw/hip screw) and 6.4mm (anti-rotation screw/hip pin). Both dynamic and static options for 4.9mm bolts were present in distal locking and the jig allowed placement of distal bolts through the jig itself. Nails were designed with option of 130 degrees and 135 degrees neck shaft angle. The dimensions of this modified short PFN is smaller than standard PFN that comes in length of 240mm, proximal diameter of 17mm, distal diameter 10-13mm and cephalic screws measuring 11mm and 6.5mm (Fig 1).
Immediate post-operative radiographs were considered as baseline for subsequent implant related measurements. Tip apex distance (TAD), quality of reduction, position of tip of lag screws in head was done on immediate post-operative radiograph. Sequential follow-up radiographs were evaluated to assess union, position of screws and to calculate the lateral slide of lag screws. Magnification of the radiographs were calculated dividing true lag screw width by screw width measured on radiograph. All lengths measured were multiplied by this factor to account for magnification. TAD was calculated by the method described by Baumgartner et al 13 and adapted to cephalomedullary nail as described by Herman et al (Fig. 2) 14 . Apex was marked in both the views for calculation of TAD. The distance between the tip of screw and apex in that particular view was defined as TAD in that view. [TADtotal = TADAP + TADlateral.] Baumgartner's original description of TAD pertains with sliding hip screw system with a large single cephalic screw. It has been extrapolated and used in cephalomedullary nails including dual screw systems such as ours 15,16 . TAD was measured for only the lag screw as hip pin gets obscured by the lag screw in lateral view. Baumgartner criteria 17 was used to assess quality of reduction. Position of the tip of lag screw in the femoral head was assessed using Cleveland zones 18 . For the measurement of lateral slide of lag screw, immediate post op and final AP radiographs were compared as described by Morihara et al 19 .
Kyle's criteria 20 was used for final functional assessment while Barthel's index 21 was used to assess level of independence in activities of daily living. IBM SPSS Statistics version 26 was used for data analysis. Apart from descriptive statistics, Wilcoxon signed-rank test and Mann-Whitney U tests were used for analysis of nonparametric data.

RESULTS
Demographic details are enlisted in Table I. Average age of the patients in our study was 71.45 years (range 30-95 years) with a median of 69.47 years. 56.66% were males while rest were females. Left and right side were affected in 40.83% and 59.17% respectively. Majority of fractures were of A2 type (68.3%). Table II enlists intra-operative details and findings. All patients were operated on a traction table and closed reduction was attempted. Only after failure of closed means (including joystick method), open reduction was done. Closed reduction was achieved in 98 patients, 14 needed joystick manoeuvre while 8 patients needed limited open reduction. Complete exposure of the fracture was not needed in any case. Predominantly 135° nail was used with 11mm diameter. There was a mean difference of 14.85mm in the sizes of the two cephalic screws used. The quality of reduction as per Baumgartner's criteria on immediate postoperation radiographs was Good in 90.83% and no patient was classified as poor reduction. Toe-touch weight bearing with walker support was immediately started postoperatively and full weight bearing was undertaken only after radiological union.
We encountered certain procedure specific intra-operative difficulties in 16 cases and have been compiled in Table III. Most frequent was difficulty in inserting 6.4mm screw which was observed in 5% of cases. Mean TADtotal was 22.8mm while TADap and TADlateral was 11.8 and 11.0mm respectively (Table IV). TAD has been poorly studied in biaxial cephalomedullary implants (having two screws) such as ours   and currently there is no proven or recommended TAD for such implants 15,16 . Most frequent position of lag screw was charted in inferior-central zone in 95 cases (Fig. 3). The next most commonly plotted position was central-central position and it was noted in 17 cases (14.16%).
Complications have been grouped and compiled in Table V. Reoperation was required in a total of 10 cases (8.3%). A total of five cases had screw cut-out and have been analysed in discussion part of the article with other cases of mechanical failure. One case was associated with deep infection. Peri-implant fracture was noted in two cases although it occurred after fracture consolidation. Isolated Z effect without loss of reduction was noted in four cases while most common complaint at final follow-up was thigh discomfort in 13 cases (10.8%). Average time to fracture 75    union was estimated to be 17.32 weeks with average shortening noted to be 4mm. Nine patients (7.5%) had a shortening of one cm or more. The overall mean lateral slide of compression screws was estimated to be 3.20mm (range 0 to 13mm), after exclusion of cases with screw failure/cut-out. Unstable fracture patterns had more slide than stable ones. A1 fractures had a mean slide of 2.30mm (0-4 mm) while A2 type had 3.42mm (0-13mm).

DISCUSSION
We searched other series for intra-operative complications and compared them in Table III 19 did not report even a single cut-out of screws (not even Z effect) leading to conclusion that anti-rotation screws being 10-15mm shorter than the lag screw prevented the cut-out. Multiple predictors of screw cut-out has been described. Using multivariate logistic regression analysis Escolar et al found TAD, suboptimal osteosynthesis and distal static locking as predictive factors for screw cut-out 32 .
Kashigar et al used univariate analysis and found TAD, calcar-TAD, Parker's ratio index and neck angle difference to be associated with screw cut-outs in cephalomedullary nails 33 . John et al studied and included biaxial cephalomedullary nails in addition to uniaxial nails. They concluded that a combination of high TAD, suboptimal position of implant and poor restoration of neck shaft angle may predispose to cut-out. However, achieving TAD within safe limits didn't appear to influence screw and device migration in dual screw nails 16 . Another technical aspect of note is the length of anti-rotation screw. PFN being a twin screw construct, the smaller screw (proximal hip pin, 6.4mm) serves the purpose of providing rotational stability while the lag screw serves load bearing function. When hip pin protrudes beyond lag screw, increased vertical forces induce Z-effect (aka Knife effect) forcing the proximal screw medially into the joint and distal lag screw to slide back laterally.
Screw cut-out rate was 5/120 in our series. One of which was associated with infection, and four without infection. We analysed TAD, position of lag screw as well as relative length of anti-rotation screw on post-operative radiographs in the screw cut-out cases. Overall TADtotal was found to be 22.8mm which was less than Fogagnolo et al 23 and Uzen et al 34 but more than Herman et al 14 (Table IV).
Calculations of mean TAD (Table VI) reflected higher values in the group with screw cut-out when compared with the one without cut-out. Mann Whitney-U test suggested that TADlat and TADtotal were significantly different in the two groups, while TADap showed a trend towards significance. Baumgartner 13 recommended TAD to be less than 25mm, albeit in a single screw construct. Some authors 23,34 have found correlation of large TAD in PFN (a double screw construct) with screw cut-outs while others 14 have refuted its use for PFN. Most of the authors however do concur with the fact that tip of lag screw must be as close to subchondral bone as possible. TAD represents both the position and depth of a screw in the femoral neck and head and was shown to be the most important predictive factor for the occurrence of a cut-out 35,36 . Geller et al 37 reported a high incidence (44%) of cut-outs in intertrochanteric fractures that were surgically fixed with a TAD of >25mm.
Ideal placement of lag screw in head is suggested to be Inferior-Central 19 . Kyuzyk et al demonstrated that biomechanical stiffness is maximised when lag screw is placed inferiorly in AP view, and central placement in lateral view maximises its load to failure 38  Both are intricately connected as angle of screw placement is inherent to the design of neck and that is prefixed, hence unless correct neck shaft angle is achieved, screw insertion may prove to be very tricky. This again emphasises on achieving as near anatomical reduction as possible.
As mentioned before, length of derotation screw has been reported as predictive factor for cut-out. Morihara et al recommended that derotation screw must be at least 10-15mm shorter than the larger lag screw 19 . Zirngibl et al 39 analysed this by drawing an imaginary line from tip of lag screw to the tip of nail and proved that anti-rotation pin protruding beyond this line had a significantly high odds ratio of 8.8 for fixation failure. They go on to suggest that, this could be the most important factor influencing the screw cut-out or cut-through rates. Analysis of relative screw lengths in femoral head revealed that in four out of five cases of screw cut-out, the anti-rotation screw was advanced either beyond the tip of lag screw or was at the same level, thus leading to increased vertical forces on the anti-rotation pin.
In short, a combination of suboptimal position of lag screw in femoral head, high TAD as well as excessively long antirotation screw were found in cases that had fixation failures. The mean operative time found in this study was lesser than that reported by Fogagnolo (83.4 min) and Morihara (77min) who used standard PFN. Some studies do quote lesser operative time 41 , however, it is unclear what constitutes operative time in studies. Whether from incision to closure or from starting of attempt at closed reduction. In our study, we included the duration of closed reduction before incision as well. Mean blood loss is considerably lesser than that occurs with standard PFN [41][42][43] .
In authors' opinion, in order to avoid screw cut-out and mechanical failure, effort needs to be directed at minimising TAD by inserting compression screw deep into the head up to 5mm below subchondral bone. In addition to ensuring adequate purchase in proximal fragment it also prevents inadvertently longer anti-rotation screws. Every effort must be directed towards careful placement of lag screw in 'safe quadrant' (inferior in AP and central in lateral view). Achieving appropriate anatomical reduction and not accepting even slight varus goes a long way in achieving this objective. Valgus reduction may be accepted, implant permitting, and may even be recommended in unstable fractures. In unstable fractures as union occurs, further impaction and varus occurs.
The overall mean lateral slide of compression screws was estimated to be 3.20mm (range 0 to 13mm), after exclusion of cases with screw failure/cut-out. This lateral slide was found out to be more in unstable fractures when compared to stable fracture patterns. A1 fractures had a mean slide of 2.30mm (0-4mm) while A2 type had 3.42mm (0-13mm). As union progresses, proximal fragment gets impacted onto distal fragment as well as the nail, leading to lateral slide of both the cephalic screws and can be a surrogate marker of collapse of the fracture. Any restriction in this lateral slide may initiate cut-out or joint penetration by the screws.
Despite these complications and mechanical failures, recovery to pre-injury functional status as per Barthel's score was found in 81.66% of the cases (change less than MCID). As per Kyle's criteria, good to excellent functional recovery was found in 88.33% (106) cases. Gadegon et al 11  In limitations, inherent to the methodology of the study which involves medical records examination, we could not use femur length as our inclusion criteria as these were not consistently mentioned in all the records. Another limitation of this study is a lack of control group.

CONCLUSION
Modified Short Proximal Femoral nail needs careful preoperative plan, followed by expert intra-operative technique coupled with good reduction. If appropriately followed, it leads to high union rate with minimal soft tissue damage. Placement of screws needs special mention and are essential for successful outcome. Safe position of screw is inferior in AP plane and central in lateral view. TAD needs to be kept to minimum. Deep insertion of lag screw into femoral head, closer to subchondral bone with a shorter anti-rotation screw which doesn't cross the tip of lag screw is equally important. Although, not devoid of complications, modified short PFN results in good functional recovery of patients with intertrochanteric fractures of femur. The shorter nail allows for easier insertion (no reaming required post isthmus) and lesser blood loss with lesser complication rates. Its shorter length renders it not suitable for fractures that extend far distal to lesser trochanter. Further studies are needed to compare the efficacy of shorter variant in Asians as well as compare it with new variant PFNA.

CONFLICT OF INTEREST
No potential conflict of interest.

FUNDING SUPPORT
None